Artificial intelligence (AI) has seemingly infiltrated many industries across the workplace, and publishing is no exception.
From plagiarism checkers to reviewers using AI tools to double check the validity of manuscript data, AI has made its way into the arena, but the real question is, what can’t AI do in publishing?
Publishers are busy focusing on different ways to integrate AI into the industry and establishing guidelines for both authors and reviewers, ranging from restricting usage altogether to outlining the specific ways it can be integrated.
However, what doesn’t seem to be a big focus just yet is the discussion on the limitations of AI and ways that it hasn’t infiltrated the industry as of 2026.
Below are a few points that show what AI can’t provide in publishing:
- Human authenticity
AI still works off human-provided prompts. Sure, you can request AI to write from a particular emotion, but it is surprising how often readers can detect the difference between something written by AI versus something written by a human being – and hint! – It’s not just the use of em dashes.
This blog by Skyline Academic highlights key differences that certain detectors look for when trying to decipher human vs. AI writing. One thing that was noted was the idea of rhythm in writing. Every human being has their own – from the long windedness of Olmstead to the shorter quips of Steinbeck, famous authors can be known for their writing style, and a lot of that has to do with rhythm! This article notes that the key differences between human and AI writing may lie in the rhythm. It states that humans “rarely write in a perfectly steady rhythm.” It then goes on to note that AI text tends to “feel very even.” Humans can’t, for the most part, be as calculated as AI, so writing styles still remain authentically human.
- Editorial Judgment
While AI can screen for syntax, word flow, or correct data reportage, AI still doesn’t have the capability to make an overall judgment on what should exist within current literature. It can make comparisons with the existing literature, but that information alone can be biased! AI tools can take in information based on a process called website scraping. This indicates that it often relies heavily on the information available on the internet, which humans know is not always based on factual accuracy! Hence, it might be making assumptions based on pre-prints that have not been verified or articles that have been uploaded that have not been reviewed through the appropriate practices. Therefore, editorial judgment is still up to humans.
While humans can be inherently biased, it is still important to have someone overseeing manuscript output based on years of experience in the field rather than prompts driven by AI. You can find more information on what AI scraping is at the link here.
AI as an Author
It is not yet common to see AI as a listed author for a manuscript. A blog titled “Should AI be attributed as an author of AI-generated works?” by Matulionyte explores the conversation behind this topic. Currently, the law behind AI and authorship is a bit murky. Matulionyte discusses a German photographer, Eldagsen, who won a photography award, but after the fact, he announced the image was generated through AI and refused to accept the award. Matulionyte goes on to talk about how this instance “provoked a public discussion on whether AI should be attributed or mentioned when a work is generated by AI or using AI.” Although AI cannot currently be listed as an author, ongoing legal developments may reshape how authorship is defined in relation to AI.
Making Meaning Beyond Metrics
AI can help create the perfect SEO tool for a story; however, often these types of searches drift users away from a publisher’s website and into more click-bait style writing that is not as verified. Forbes reports that “AI search engines send 96% less referral traffic to news sites and blogs than traditional Google search.” As a result, publishers cannot rely on metrics to understand what drives users to their sites.
While data is one thing, meaning is another. Publishers help bring out why a story matters. AI might be able to help optimize clicks or engagement, but the broader purpose in publishing is why a story matters, how the information can contribute to a broader conversation, and thinking in context of who it serves, which, currently, is still humans!
With all the talk about AI and its infiltration into publishing, it is important to take a step back and remember what still makes publishing human. The connections made between content and storytelling are still very much alive and don’t seem to be going anywhere anytime soon. Ultimately, while technology continues to evolve, the heart of publishing remains rooted in creativity, authenticity, and connectivity!
By Emma Montes
